From: Atish Patra <atish.patra@linux.dev>
To: Will Deacon <will@kernel.org>,
Atish Patra <atish.patra@linux.dev>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Anup Patel <anup@brainfault.org>,
Ian Rogers <irogers@google.com>, Rob Herring <robh@kernel.org>,
Paul Walmsley <pjw@kernel.org>,
Mark Rutland <mark.rutland@arm.com>,
Jiri Olsa <jolsa@kernel.org>,
Krzysztof Kozlowski <krzk+dt@kernel.org>,
James Clark <james.clark@linaro.org>,
Namhyung Kim <namhyung@kernel.org>
Cc: linux-riscv@lists.infradead.org,
linux-arm-kernel@lists.infradead.org,
Conor Dooley <conor@kernel.org>,
linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
devicetree@vger.kernel.org
Subject: [PATCH v9 05/20] RISC-V: Define indirect CSR access helpers
Date: Fri, 07 Aug 2026 01:09:00 -0700 [thread overview]
Message-ID: <20260807-counter_delegation-v9-5-58658104e487@meta.com> (raw)
In-Reply-To: <20260807-counter_delegation-v9-0-58658104e487@meta.com>
From: Atish Patra <atishp@rivosinc.com>
The indirect CSR requires multiple instructions to read/write CSR.
Add a few helper macros for ease of usage.
These have to be macros rather than functions. csr_read()/csr_write()
stringify their CSR argument into the inline asm template via
__ASM_STR(), so the CSR number must be a literal token; passing it as a
function parameter emits "csrr %0, iregcsr", which the assembler rejects
with "unknown CSR `iregcsr'". The stringification happens in the
preprocessor, before inlining or constant propagation, so it cannot be
worked around by forcing inlining or by only ever passing constants -
gcc 12, gcc 16 and clang 22 all reject it alike. Underneath, csrr/csrw
encode the CSR as a 12-bit immediate and RISC-V has no register-indirect
form, which is also why asm/csr.h keeps every one of its accessors as a
macro.
Signed-off-by: Atish Patra <atishp@rivosinc.com>
Reviewed-by: Charlie Jenkins <thecharlesjenkins@gmail.com>
Tested-by: Charlie Jenkins <thecharlesjenkins@gmail.com>
Link: https://patch.msgid.link/20260701-counter_delegation-v8-5-7909f863a645@meta.com
[pjw@kernel.org: expand "ind" abbreviation]
Signed-off-by: Paul Walmsley <pjw@kernel.org>
---
arch/riscv/include/asm/csr_indirect.h | 50 +++++++++++++++++++++++++++++++++++
1 file changed, 50 insertions(+)
diff --git a/arch/riscv/include/asm/csr_indirect.h b/arch/riscv/include/asm/csr_indirect.h
new file mode 100644
index 000000000000..0f558fac8f5f
--- /dev/null
+++ b/arch/riscv/include/asm/csr_indirect.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
+#ifndef _ASM_RISCV_CSR_INDIRECT_H
+#define _ASM_RISCV_CSR_INDIRECT_H
+
+#include <linux/irqflags.h>
+
+#include <asm/csr.h>
+
+/*
+ * These have to be macros rather than functions: csr_read()/csr_write()
+ * stringify their CSR argument into the inline asm template via __ASM_STR(),
+ * so the CSR number must be a literal token at preprocessing time. Passing it
+ * as a function parameter emits "csrr %0, iregcsr", which no assembler can
+ * resolve. RISC-V has no register-indirect form of csrr/csrw - the CSR is a
+ * 12-bit immediate - so the sireg CSR selecting the indirect window cannot
+ * itself be a variable.
+ */
+#define csr_indirect_read(iregcsr, iselbase, iseloff) ({ \
+ unsigned long __value = 0; \
+ unsigned long __flags; \
+ local_irq_save(__flags); \
+ csr_write(CSR_ISELECT, (iselbase) + (iseloff)); \
+ __value = csr_read(iregcsr); \
+ local_irq_restore(__flags); \
+ __value; \
+})
+
+#define csr_indirect_write(iregcsr, iselbase, iseloff, value) ({ \
+ unsigned long __flags; \
+ local_irq_save(__flags); \
+ csr_write(CSR_ISELECT, (iselbase) + (iseloff)); \
+ csr_write(iregcsr, (value)); \
+ local_irq_restore(__flags); \
+})
+
+#define csr_indirect_warl(iregcsr, iselbase, iseloff, warl_val) ({ \
+ unsigned long __old_val = 0, __value = 0; \
+ unsigned long __flags; \
+ local_irq_save(__flags); \
+ csr_write(CSR_ISELECT, (iselbase) + (iseloff)); \
+ __old_val = csr_read(iregcsr); \
+ csr_write(iregcsr, (warl_val)); \
+ __value = csr_read(iregcsr); \
+ csr_write(iregcsr, __old_val); \
+ local_irq_restore(__flags); \
+ __value; \
+})
+
+#endif
--
2.53.0-Meta
next prev parent reply other threads:[~2026-08-07 8:09 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-07 8:08 [PATCH v9 00/20] Add Counter delegation ISA extension support Atish Patra
2026-08-07 8:08 ` [PATCH v9 01/20] RISC-V: perf: fix resource cleanup on driver probe failure Atish Patra
2026-08-07 8:30 ` sashiko-bot
2026-08-07 8:08 ` [PATCH v9 02/20] RISC-V: Add Smcsrind and Sscsrind ISA extension CSR definitions Atish Patra
2026-08-07 8:08 ` [PATCH v9 03/20] RISC-V: Add Smcsrind and Sscsrind ISA extension definition and parsing Atish Patra
2026-08-07 8:08 ` [PATCH v9 04/20] dt-bindings: riscv: add Smcsrind and Sscsrind ISA extension descriptions Atish Patra
2026-08-07 8:09 ` Atish Patra [this message]
2026-08-07 8:27 ` [PATCH v9 05/20] RISC-V: Define indirect CSR access helpers sashiko-bot
2026-08-07 8:09 ` [PATCH v9 06/20] RISC-V: Add Smcntrpmf extension parsing Atish Patra
2026-08-07 8:09 ` [PATCH v9 07/20] dt-bindings: riscv: add Smcntrpmf ISA extension description Atish Patra
2026-08-07 8:09 ` [PATCH v9 08/20] RISC-V: Add Ssccfg extension CSR definition Atish Patra
2026-08-07 8:09 ` [PATCH v9 09/20] RISC-V: Add Ssccfg/Smcdeleg ISA extension definition and parsing Atish Patra
2026-08-07 8:25 ` sashiko-bot
2026-08-07 8:09 ` [PATCH v9 10/20] dt-bindings: riscv: add Counter delegation ISA extensions description Atish Patra
2026-08-07 8:09 ` [PATCH v9 11/20] RISC-V: perf: Restructure the SBI PMU code Atish Patra
2026-08-07 8:09 ` [PATCH v9 12/20] RISC-V: perf: Modify the counter discovery mechanism Atish Patra
2026-08-07 8:29 ` sashiko-bot
2026-08-07 8:09 ` [PATCH v9 13/20] RISC-V: perf: Add a mechanism to defined legacy event encoding Atish Patra
2026-08-07 8:09 ` [PATCH v9 14/20] RISC-V: perf: Implement supervisor counter delegation support Atish Patra
2026-08-07 8:34 ` sashiko-bot
2026-08-07 8:09 ` [PATCH v9 15/20] RISC-V: perf: Skip PMU SBI extension when not implemented Atish Patra
2026-08-07 8:09 ` [PATCH v9 16/20] RISC-V: perf: Use config2/vendor table for event to counter mapping Atish Patra
2026-08-07 8:43 ` sashiko-bot
2026-08-07 8:09 ` [PATCH v9 17/20] RISC-V: perf: Add legacy event encodings via sysfs Atish Patra
2026-08-07 8:40 ` sashiko-bot
2026-08-07 8:09 ` [PATCH v9 18/20] RISC-V: perf: Add Qemu virt machine events Atish Patra
2026-08-07 8:41 ` sashiko-bot
2026-08-07 8:09 ` [PATCH v9 19/20] tools/perf: Support event code for arch standard events Atish Patra
2026-08-07 8:09 ` [PATCH v9 20/20] tools/perf: Add RISC-V CounterIDMask event field Atish Patra
2026-08-08 17:36 ` [PATCH v9 00/20] Add Counter delegation ISA extension support Paul Walmsley
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260807-counter_delegation-v9-5-58658104e487@meta.com \
--to=atish.patra@linux.dev \
--cc=acme@kernel.org \
--cc=anup@brainfault.org \
--cc=conor@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=irogers@google.com \
--cc=james.clark@linaro.org \
--cc=jolsa@kernel.org \
--cc=krzk+dt@kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=linux-riscv@lists.infradead.org \
--cc=mark.rutland@arm.com \
--cc=namhyung@kernel.org \
--cc=pjw@kernel.org \
--cc=robh@kernel.org \
--cc=will@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox